Manuzio: A Model for Digital Annotated Text and Its Query/Programming Language

Abstract

More and more large repositories of texts which must be automatically processed represent their content through the use of descriptive markup languages. This method has been diffused by the availability of widely adopted standards like SGML and, later, XML, which made possible the definition of specific formats for many kinds of text, from literary texts (TEI) to web pages (XHTML). The markup approach has, however, several noteworthy shortcomings. First, we can encode easily only texts with a strict hierarchical structure while text has often concurrent hierarchies. Then, extra-textual information, like metadata or annotations, can be tied only to the same structure of the text and must be expressed as strings of the markup language. Third, queries and programs for the retrieval and processing of text must be expressed in terms of languages like XQuery [4], in which every document is represented as a tree of nodes; for this reason, in documents where parallel, overlapping structures exists, the complexity of XQuery programs becomes significantly higher.Consider, for instance, a collection of classical lyrics, with the two parallel hierarchies lyric > stanzas > verses > words, and lyric > sentences > words, with title and information about the author for each lyric, and where the text is annotated both with commentary made by different scholars, and with grammatical categories in form of tree-structured data. Such a collection, if represented with markup techniques, would be very complex to create, manage and use, even with sophisticated tools, requiring the development of complex ad-hoc software.To overcome the some of the above limitations partial solutions exist (see for instance [3]), but at the expense of greatly increasing the complexity of the representation through difficult to read markup extensions, like the so-called "milestone" elements. Moreover, markup query languages need to be extended to take these solutions into consideration [1], making even more difficult to access and use such textual collections.In the project "Musisque Deoque. A digital archive of Latin poetry, from its origins to the Italian Renaissance" sponsored by the Italian MIUR, we have built a model and a language to represent repositories of literary texts with any kind of structure, with multiple and scalable annotations, not limited to textual data,

Crossing Experiences in Digital Epigraphy

“…This substantial rewriting of the project source code has been the occasion to address, in a more formal manner, problems of data persistency by formalizing a new conceptual schema. This new schema has been deeply influenced by the high level of formalism of the Manuzio project (Maurizio & Orsini, 2010a), from which Sinleqiunnini differs in the structure of implementation (Maurizio & Orsini, 2010b).…”